Do CBT-I effects on insomnia severity vary by chronic disease type?

How large is the improvement, and does it hold across conditions?

Across 67 randomized trials involving 5,232 adults with chronic disease, CBT-I produced a large improvement in insomnia severity, with an effect size of 0.98 [1]. In plain terms, an effect size near 1 means the average treated person improved by about one standard deviation—a substantial shift in symptoms, comparable to the effects seen in people without chronic disease [1]. Sleep efficiency and the time it takes to fall asleep also improved, by moderate amounts [1].

The key question—whether the benefit depends on the disease—was tested directly. Subgroup analyses found that efficacy was similar across a range of chronic disease subgroups, including cancer, chronic pain, irritable bowel syndrome, and stroke [1]. That pattern is reinforced by narrower reviews: in cardiovascular disease, CBT-I effectively reduced insomnia symptom severity and shortened sleep onset [3], and in chronic obstructive pulmonary disease, a randomized trial found a significant effect of CBT-I on insomnia severity [2].

Why the disease label may matter less than the insomnia itself

CBT-I targets the behaviors and thoughts that keep insomnia going—such as irregular sleep schedules, spending too long in bed awake, and anxious beliefs about sleep—rather than the underlying medical condition. If those maintaining factors are similar across diseases, the treatment should work similarly, which is consistent with the similar subgroup results [1].

In the COPD trial, the effect of CBT-I on insomnia severity appeared to operate through changes in beliefs about sleep, suggesting that shifting unhelpful sleep-related cognitions is part of how the therapy helps in at least some chronic disease populations [2]. That is a proposed mechanism, not proof that every disease works the same way.

What this means for someone with a chronic disease and insomnia

The practical takeaway is that a chronic disease diagnosis is not a reason to expect CBT-I to work less well. Satisfaction was high and dropout averaged 13.3%, with treatment-related adverse effects rare [1]. Longer treatment was associated with better outcomes for sleep efficiency and sleep onset latency, suggesting that giving the therapy enough time matters [1].

One caveat sits beside these results: the evidence base is uneven across diseases. The large meta-analysis covers cancer, chronic pain, irritable bowel syndrome, and stroke among others [1], while dedicated trials exist for cardiovascular disease [3] and COPD [2]. For conditions not well represented, the similar-across-subgroups finding is encouraging but less directly tested. The authors of the large review call for research into treatment adaptations for specific populations and better access to CBT-I in medical settings [1].

Sources used in this answer

Related research questions